Author(s):
Henriques, Roberto André Pereira
Date: 2006
Persistent ID: http://hdl.handle.net/10362/3641
Origin: Repositório Institucional da UNL
Subject(s): Cartograms; Neural networks; Kohonen self-organizing maps; Geographic information systems; Population; Matlab; Cartogramas; Redes neuronais; Mapas auto-organizáveis de Kohonen; Sistema de informação geográfica; População; Matlab
Description
Dissertação apresentada como requisito parcial para obtenção do grau de Mestre em Ciência e Sistemas de Informação Geográfica
The basic idea of a cartogram is to distort a map. This distortion comes from the substitution of area for some other variable (in most examples population). The objective is to scale each region according to the value it represents for the new variable, while keeping the map recognizable. The first cartograms were created to show the geographic distribution of population, in the context of human geography. Cartograms can be seen as variants of a map. The difference between a map and a cartogram is the variable that defines the size of the regions. In a map this variable is the geographic area of the regions, while in the cartogram any other georeferenced variable may be used. In this dissertation we present a general method for constructing density-equalizing projections or cartograms, using the basic SOM algorithm, providing a tool for geographic data presentation and analysis.